Download Dr Phillip Toner - Productivity Commission

‘An Economic Critique of the Case for Contracting-out Government Funded Human
Services’
A Submission to the Productivity Commission Inquiry into Reforms to Human Services
by
Dr Phillip Toner
Honorary, Department of Political Economy, University of Sydney
Contents
1. Introduction
1.1 Broader context for the PC Inquiry Into Human Services- Neoliberal public policy
2. Problems with competition in the delivery of human services
2.1 Benefits of competition
2.2 Transaction cost economics
2.2.1 Incentives and opportunism
2.3 Competition, contracting-out and innovation
2.4 How will competition work in practice?
3. Impossibility of precise cost comparisons and cost-benefit analysis in heterogeneous
human services
3.1 Establishing a causal relation between services and outcomes
3.2 What is an ‘efficient cost’?
3.3 Problems in Comparing Prison Cost and Performance
3.3.1 Cost per prisoner per day
3.3.2 Selection of the four independent variables to adjust benchmark targets
1
1. Introduction
This submission focuses on just 2 elements of the Productivity Commission’s Terms of
Reference; these are the application of competition to the contracting-out of human services
and the difficulties in creating, monitoring and regulating efficient contracts between
governments and for-profit or not-for- profit non government entities for the delivery of these
services. The study uses the contracting-out of prison services in Western Australia as a case
study to demonstrate the theoretical and practical problems of efficient contracting out.
Specifically, it draws on the recent reports by the WA Economic Regulation Authority
(ERA), an entity similar to the PC in terms of its scope of work and adherence to the
neoclassical approach to economic analysis and public policy, which strongly advocated for a
extension of the existing limited contracting-out of WA prisons. 1
By selecting an appropriate economic theory, especially a simplified version of neoclassical
analysis, the case for contracting out government funded human services can be readily
made. This is due to assumptions regarding perfect information, absence of perverse
incentives and complete contracting. It is imperative however, that the PC consider the
outcomes of actual contracting out of government services, because in many instances the
actual outcomes do not meet the high expectations of proponents. This gap between high
expectations and reality is due to fundamental flaws in the original economic case for
contracting-out. This is indeed the result in the specific instance of the contracting out of WA
prisons.
1.1 Broader context for the PC Inquiry Into Human Services- Neoliberal public policy
The neoliberal approach to public policy has dominated western, and primarily Anglophone,
states over the last 3-4 decades. In Australia the key pivot point in public policy was the 1993
Hilmer Report into National Competition Policy, which copied the neoliberal public policy
template from the UK and established the vast legislative and bureaucratic apparatus in
federal and state governments, such as COAG, the ACCC and ERA, for its implementation.
All major political parties, with minor variations, subscribe to the neoliberal doctrine.2 The
recent Federal Competition Policy Review (Harper Review, 2015) recapitulated the
arguments in Hilmer and advocated inter alia for extending contracting-out to government
funded ‘human services’.
The following brief account focuses on three core aspects of neoliberalism and describes how
they inform the work of ERA and the PC. These ideas are reducing the role of the state in the
This submission is based on reports undertaken for the Western Australian Prison Officers’ Union (WAPOU)
in response to the WA Economic Regulation Authority’s (2015) Inquiry into the Efficiency and Performance of
Western Australian Prisons
(https://www.erawa.com.au/cproot/13841/2/WAPOU%20Submission%20to%20Prisons%20Draft%20Report.pd
f)
2
Wolfgang Streeck (2015: 75) provides this succinct summary of a few key features of neoliberalism: ‘In short,
it is the idea, and the corresponding practice, that it is the duty of governments to install, impose and enforce
market relations wherever possible. Political intervention in the economy must refrain from redistribution,
except from the bottom to the top, to create ‘incentives’ for growth. The assumption is that people at the top
work more if they pay lower taxes on higher incomes, whereas people at the bottom work more if you take away
their social security and lower their wages. It also means that public provision should be replaced with private
purchases, so people work more. Whereas under Keynesianism economic growth was expected to come from
the allocation of spending power to people with low incomes, under today’s neo-liberalism...growth is to be the
result of income flowing to the already rich.’ For a detailed account of the historical development of
neoliberalism see Cahill et al (2012).
1
2
delivery of publicly funded services through means such as privatisation and contracting out;
adherence to ‘managerialism’ and the centrality of financial incentives in driving
performance improvement. This section also briefly highlights some of the adverse economic
and social outcomes from the adoption of neoliberal policies.
Neoliberalism is a set of economic, political and ethical propositions. The core ideas are
based on the pre-Keynesian theory of economic laissez-faire, or more formally a very
simplified version of neoclassical economics, to provide a rationale or justification for its
approach. Neoclassicals view the economy as a self-regulating mechanism in which
unfettered competition in production, labour and capital markets is the key to prosperity. This
view supports reducing the existing multiple roles of the state in the economy. It supports
reducing the share of total state expenditure and tax in the economy; reducing its role in the
direct provision of services such as utilities, welfare, education and prisons and supports
deregulation of economic activity. Aside from these economic arguments, reducing the role
of the state is also supported by neoliberal political analysis of the state. It views the state as
irredeemably captured by ‘rent seeking’ interests in industry and society, including politicians
and bureaucrats, who use the taxing and regulatory power of the state for self-serving ends.
Neoliberalism also views society as essentially a self-organising organism wherein, following
Adam Smith, the pursuit of individual self-interest is transformed into an overall harmony of
interests across society.
Central to neoliberal public policy is substantially reducing the ‘size of the state’ or, where
state activities are in the nature of a ‘public good’, have them subject to competition for their
delivery. Public goods are economic activities that are essential for the operation of a modern
economy but which, due to factors such as externalities and non-excludability, will not be
supplied, or be under-supplied, by a private market. These activities include services like
education, defence and law and order. Other services, especially infrastructure provision like
electricity, telecommunications, rail and roads are ‘natural monopolies’ in that generally it is
wasteful to duplicate the infrastructure by having competing private providers. Prior to the
rise of neoliberalism state provision of infrastructure services was justified to prevent private
owners exploiting monopoly pricing power; the state generally has much lower borrowing
costs to fund their construction than the private sector and there are often compelling reasons
for parts of these services to be subsidised. Examples of the latter are uniform price of stamps
for letters to ensure communication for citizens within a country or subsidising passenger rail
services to reduce car pollution and accidents.
As noted above public goods must be provided but, according to neoliberal public policy,
they need not be supplied by the state directly. There are three neoliberal solutions:
corporatisation of public services, whereby the broad objectives of a publicly owned business
are reduced to a single goal of maximising return on assets; privatisation and contracting-out
of a publicly funded entity.
These solutions rely on the concept of ‘competition’ and the accompanying benefits of choice
and efficiency. Competition is both a means and an end in neoliberal public policy.
Neoliberalism strongly endorses the idea that the singular focus of the firm must be
maximising of ‘shareholder value’, or more formally profit maximisation. Competition is
argued to promote allocative efficiency (production is optimally organised to meet the needs
of consumers) and productive efficiency (marginal revenue from factors equals marginal
output of factors). Applied to the WA prison system ERA (2015: 251) suggests ‘competition
3
encourages businesses to compete for customers (in this case the Department) and can result
in lower prices, better quality, greater choice, and higher levels of innovation’.
The central idea of managerialism is managerial prerogative. John Quiggin (2003) has
observed: ‘the central doctrine of managerialism is that the differences between such
organisations as, for example, a university and a motor-vehicle company, are less important
than the similarities, and that the performance of all organisations can be optimised by the
application of generic management skills and theory. It follows that the crucial element of
institutional reform is the removal of obstacles to ‘the right to manage’...The rise of
managerialism has gone hand in hand with that of the radical program of market-oriented
reforms variously referred to as Thatcherism, economic rationalism and neoliberalism... it is
claimed the optimal policy is to design organisations that respond directly to consumer
demand, and to operate such institutions using the generic management techniques applicable
to corporations of all kind. The main features of managerialist policy are incessant
organisational restructuring, sharpening of incentives, and expansion in the number, power
and remuneration of senior managers, with a corresponding downgrading of the role of
skilled workers, and particularly of professionals’.
To get public and private operators of prisons to conform to the needs of the customer
financial incentives and dis-incentives are necessary. ‘Central to establishing an efficient
prison system is the identification of the incentives that align the interests of the prison
operator with those of the State....Incentives are central to a well-designed prison system
because, if they are harnessed appropriately, they will maximise the chance that prisons
achieve the objectives set out in the performance framework. Incentives can encourage
prisons to find more innovative ways to effectively rehabilitate prisoners and reduce costs,
ultimately improving performance’ (ERA 2014: 49). (The very considerable problems of
constructing contracts and incentive systems that do not lead to perverse effects and
opportunism on the part of contractors and the cost of monitoring performance are dealt with
below).
Based on the actual outcomes of neoliberal policies a different view emerges of the neoliberal
public policy agenda. Neoliberal policies do not reduce the role of the state in the economy;
rather they change its form. For example, contracting-out publicly funded, but privately
delivered services, does not reduce public expenditure, it simply expands the scope for capital
(private profit making firms) to enter new spheres of economic activity from which it was
previously excluded. Frequently contracting-out public services has led to failure of
provision, cost blow-outs causing serious economic and social problems and blurring
responsibilities for outcomes between the contractor and contractee. For example, the threat
of financial loss or ‘abatements’ for failure to meet contractual obligations was insufficient
incentive for Serco to maintain standards in surgical instrument sterilisation and cleaning
functions at Fiona Stanley hospital (ABC June 2015). Similarly, the threat of loss of the WA
prison transport contract was insufficient for Serco to meet contractual obligations (The
Australian 2015). Other contracted-out activities, such as employment and job seeker services
have been ‘rorted’ tens of millions of dollars by private and not for profit providers (ABC
February 2015).One of the most stark examples of the failure of contracting-out has been the
waste of hundreds of millions of dollars of public and private money on publicly funded
private vocational education and training colleges (Bita 2015; Toner 2014).
Similarly, privatisation does not eliminate the state but shifts its role to that primarily of a
regulator of private monopolies or oligopolies. For a range of reasons the outcomes of this
4
regulation are very often, very poor. For example, following privatisation and corporatisation
the electricity sector, it became one of the most heavily regulated under National Competition
Policy rules. However, because, rather than despite of this regulation, between 2007 and 2012
the weighted average Consumer Price Index for household electricity prices across all capital
cities increased by 57%, compared to an overall increase in the CPI of just 14.5%. Average
weekly earnings increased by 27.5% over the same period (Chester 2013: 3). The reasons for
these bad outcomes include the inability of regulators to know the true costs and profitability
of regulated private firms; perverse incentives created by the deeply flawed system for
regulating prices, such as the guaranteed return on investment in the electricity network, and
the need to provide high returns to encourage both initial sale of the assets and bidding for
future privatisations.3
3
In support of its advocacy of greater competition ERA (2015: 237) cites the recent Federal Competition Policy
Review (Harper Review) which promoted ‘the need for greater competition in the human services sector’, like
disability services, health and education ’which has traditionally been shielded from competition, to improve
performance’. The economics editor for the Fairfax group of newspapers, Ross Gittins wrote, more in weariness
than in anger, that ' to frame them [human services] as part of competition policy is an old economists’ trick:
take an area that has always been outside the market-place and marketise it. Take the world as it is and make it
more like the textbook assumes it to be. Apply the economists’ two magic answers: getting the incentives right
and introducing competition and choice and everything will fix itself without the economist ever needing to
come to grips with the causes of the particular inefficiencies that are causing the problem. Brilliant but often
disastrous’. He then goes on to cite the multiple failures in contracting-out child care, TAFE, job services, and
public funding of private education. ‘Think of all the money federal governments have pumped into private
schools in the sacred name of choice, without any evidence of this wider competition leading to higher standards
of education on either side of the fence’ (Ross Gittins 30/11/2015, SMH, p. 28).
5
2. Problems with competition in the delivery of human services
2.1 Benefits of competition
ERA identifies four benefits from contracting-out and competition in the provision of
publicly funded prisons- choice; better quality; innovation and reduced cost. ‘Public money is
scarce, and the Government is responsible for ensuring that these funds are spent
appropriately. The Department cannot be sure that resources are being spent efficiently if it
does not consider all available options. It is through robust competition that the Department is
able to consider the options that are available to it’ (ERA 2015: 250). This section focuses on
just two of these claims, that prisons satisfy the conditions necessary for efficient contractingout to achieve competition between suppliers and that completion promotes innovation in
prison services. The issue of quality and cost is dealt with in section 3.
2.2 Transaction cost economics
The argument that contracting-out is the optimal approach to efficiency is, frankly, silly. First
there are many ways an organisation can seek to ensure its efficiency, aside from contractingout its operations. For example, it can identify global best practice organisations that are
widely recognised as high performing and seek to emulate their practices. For prisons this
would include high performance on security and low recidivism. An organisation can
thoroughly evaluate its practices to determine what are producing desired outcomes, what do
not, the reasons for this outcome and effect remedial change. Third, it can investigate whether
the resourcing for the task it has set itself is adequate.
ERA’s argument for contracting-out and competition is founded on a simple and indeed,
simplistic, economic theory and a priori judgement that competition is an optimal method for
resource allocation. The decision by an organisation to contract-out activities and the
conditions for optimal contracting have been the subject of intense academic research, a field
known as Transaction Cost Economics (TCE). TCE teaches us, and leading orthodox
economists have widely accepted that determining how an organisation should allocate its
means to achieve its desired ends should be founded on a detailed empirical analysis of costs,
risks and potential benefits.
TCE finds that direct administrative guidance is the most appropriate form of governance
where the following conditions are met:
 the service or product is a core activity and central to the commercial survival and/or role
of the organisation. Under these conditions contracting-out introduces risk both the
organisation and to the wider society, if failure to deliver an adequate service, generates
negative social externalities. Failure to deliver an adequate quality and quantity of prison
services unarguably generates these wider social costs. The significance of activities
generating large social costs arising from contracting-out to a private provider is that these
costs are not borne by a private contractor. The contractor therefore, has no commercial
interest in their mitigation
 the activity is subject to significant uncertainty regarding the volume of services to be
delivered and/or its specification demanding therefore flexibility and rapid response to
changed circumstances
 it is difficult to precisely identify all inputs, outcomes and their precise costs entering into
production of the service making it difficult or impossible to create an efficient contract
6
 related to this latter problem is information asymmetry between a contractor and
contractee. Where a contractee knows more about the contracted service than they declare
this can result in exploitation of a contractee, especially when they seek to shift risk from
themselves onto the contractee. The same problem can happen in reverse.
These conditions describe the WA prison system. TCE highlights the risk to both contractee
and contractor arising from these conditions. ERA just does not engage with the risks
associated with contracting out prison operations.4
Unfortunately, for ERA a particularly pertinent example of exactly the risks from
contracting-out described by TCE is given in the Serco plc Annual Report 2014-15. Serco is a
large multinational service delivery organisation, prominent in the market for contracted out
government services. Serco is the contractee for two prisons in WA. Over the 2014-15
financial year Serco brought losses of £632 pounds to account on revenue of £3.9bn. These
realised losses and contingent liabilities not yet brought to book are described as being
‘onerous’. The Annual Report attributes these losses in part to opportunism/malfeasance on
the part of some employees, such as the UK Electronic Prisoner Monitoring Programme
which resulted in contract cancellation and large damages payments. More importantly, Serco
simply made errors assessing costs and risks before entering into many contracts. The
combined effect of realised losses, future anticipated losses, failure to deliver current
contracted services and wider ‘reputational damage’ is to imperil survival of the firm.
Confirming a prediction of TCE, the broader social costs of failure to deliver core
government services, which Serco describes as its core business, is barely if ever noted in the
Annual Report. The overwhelming concern is diminished shareholder value.
It is worth quoting the Annual Report at some length:
‘Our success depends on our ability to write contracts which balance risk and reward and
meet the contractual requirements into which we have entered with our customers, which
could be through direct delivery of services, through the use of sub-contractors, or through
Joint Venture consortium partners. We are subject to risks associated with bidding for and
entering into contracts (most of which are multi-year and/or fixed price contracts), including
correctly assessing and agreeing pricing terms that provide for a level of return on the
contract appropriate to the risks involved, accurately anticipating the costs of strict
performance conditions, employee requirements and other obligations, correctly evaluating
contractual and operational risks, and the risks of potential early termination or change of
scope of contracts by customers...Unclear, ambiguous, misread, misinterpreted contract
obligations and expectations of contract performance can result in perceived or actual
contract non-compliance and/or poor performance.
We are party to a number of contracts that are multi-year, fixed price, carry strict
performance conditions and/or contain volumetric or other risks relating to original bid
assumptions that have proven incorrect, and we expect to result in losses, as a result of which
we have determined the contracts to be onerous. In the second half of 2014 there were several
ERA (2015: 267-271) devotes several pages to outlining the principles ‘of good contract management’. But
these are dealt at a very high level of generality, taken as they are from a manual by the Australian National
Audit Office (Developing and Managing Contracts – Better Practice Guide, Canberra, Government of
Australia, 2012). The ANAO describes the purpose of the manual in the Forward. ‘The Guide is intended to be a
general reference document for senior managers, contract managers and stakeholders who are involved in the
development and management of contracts. The Guide does not address specific issues that relate to high-value,
complex contracts’. This surely describes contracting-out a prison.
4
7
contracts where operational issues and/or discussions with our customers resulted in us
substantially revising upwards our estimates of the costs to complete our obligations under
such contracts or lowering our revenue expectations. A risk-based independent review of our
principal contracts to identify loss-making contracts against a specific scope revealed that we
have a number of contracts that have, or are expected to result in, or could result in material
loss, which we have determined to be onerous. The costs to complete these contracts
outweigh the financial benefit, and they are, therefore loss-making resulting in lower than
expected returns and economic damage for which provisions have been made in the accounts,
and there is a risk that the losses damage our reputation...It has not been practical to complete
a full legal, operational and financial review of every contract, given the scale, complexity
and volume of the contracts and the cost and time that this would have taken. No assurance
can be given that the onerous provisions that we have recorded will be sufficient to cover the
losses ultimately incurred under the contracts for which onerous provisions have been made
or that further provisions for such contracts will not be required in the future or that the costs
of fulfilling other contracts to which any member of the Serco Group is a party will not
exceed the actual or expected economic benefit under such contracts resulting in the need for
further onerous provisions for such contracts. Inevitably, the review of contracts was carried
out at a specific point in time and with the information available at that time, which may not
prove to have been entirely accurate or complete. Further, the review could not cover all
possible circumstances on all contracts under which losses could in the future possibly be
incurred. Contracts that have not been reviewed may in future become loss-making; and
losses on contracts that have been reviewed may turn out to be worse if, for example, the
review was based on information which is subsequently superseded or revised in light of any
further review work undertaken or circumstances under the contract change. Similarly, we
may have over-estimated the provisions taken with respect to one or more of our contracts.
The onerous provisions that have been made are management's best judgement at the time of
the review. The onerous provisions are subject to change if additional information comes to
light in the future. If additional provisions and/or increased costs need to be recognised in the
future, this may result in lower returns and economic, reputational and other impacts
associated with onerous contracts, which could materially adversely affect our business,
financial condition, results of operations and prospects...If the proposed Rights Issue does not
proceed and we are unable to obtain further waivers of our financial covenants under our
financing agreements, and we are unable to avoid a breach of our financial covenants or
cross-defaults through the successful implementation of one or more funding alternatives
including proposed disposals, shareholders are at risk of losing all or a substantial amount of
their investment in the Group and the Group is at risk of not being able to continue as a
going concern’ (SERCO 2015: 15-16, italics added).
2.2.1 Incentives and opportunism
ERA provides an unbalanced analysis of economic incentives and the profit motive as it
presents these as producing primarily positive outcomes for prison performance and does not
give sufficient weight to the many examples where financial incentives have either promoted
opportunism on the part of prison contractors or have been unable to prevent unethical and/or
criminal behaviour. ERA deals with these issues at a high level of generality with anaemic
statements like ‘the performance measure must not encourage perverse behaviour’ (ERA
2015: 7). It does not enter into a detailed analysis of the problems of contracting-out prison
services and certainly fails to consider even the fundamental question, is this essential public
service even suitable for contracting-out?
8
At its base ERA’s argument that contracting-out will lead private providers to deliver
efficient, effective and ethical services rests on the positive incentive, the promise of profit,
and negative incentives, financial loss for failure to meet performance standards and/or
contract termination and possible damage to the contractee’s goodwill.5 For example,
‘contracts...establish a range of financial incentives and penalties that can be applied to Serco
depending upon its performance. Having the outcomes clearly articulated in contracts, and
subject to specific financial incentives and penalties, makes Serco accountable for achieving
the outcomes expected by the Department...In addition, Serco is subject to fixed term
contracts and knows that, if it does not meet the required standards, its contract will not be
renewed. This would not only result in the immediate loss of business, but also affect its
reputation and potentially its chances of winning business in other jurisdictions’ (ERA 2015:
119).6
It s difficult to reconcile this high-minded understanding of contracting with the following
four brief examples of contrary behaviour:
 in June 2015 DoCS chose not to renew Serco’s prison transport contract for another five
years apparently due to a series of high profile escapes (http://www.abc.net.au/news/201506-16/wa-prisoner-transport-contract-to-go-to-tender/6550474)
 Serco (2015: 6) records a number of ‘traumatic events of 2013’ that occurred in the UK,
specifically ‘overbilling in our [prisoner]Electronic Monitoring contracts and misreporting
of data on the Prisoner Escort & Custody Services’ (p.6). This resulted in the cancellation
of contracts; penalties of around £70m; ‘investigation by the Serious Fraud Office’ and
further potential ‘financial penalties and mandatory debarment from prequalifying for
future contracts with UK Government entities’ (Serco 2015: 17)
 Serco had its contract to run New Zealand’s Mt Eden prison cancelled and faces penalties
of $1m after a number of severe contractual failures
(http://m.nzherald.co.nz/nz/news/article.cfm?c_id=1&objectid=11558364)7
 the possible problems for clarity and transparency that can arise from related party
transactions have already been mentioned in relation to the Private Registered Training
Organisation ASPAC and Acacia. Another important example is the training of prison
guards for Serco’s two privately run New Zealand prisons by a Private Registered
5
It hardly needs commenting but all of these benefits and costs were articulated in the Serco 2014-15 Serco plc
Annual Report, though it laments they did not prevent all of the risks of non-performance being realised and
imperilling the company’s survival.
6
ERA's argument for the benefits of fixed term contracts can easily be turned on its head. A finite term of a
contract can encourage a contractor to maximise its return given that contract renewal can be unrelated to
performance. This is especially in a policy field as contested as contracting-out prisons. It will be recalled that
Serco noted that continuity of its business model was heavily dependent on government goodwill to the general
policy of contracting-out. In addition, managers of contracted-out services can be subject to personal incentives
or bonuses from head office, such as those requiring reduced costs, that can imperil meeting contract conditions.
7
The Mt Eden debacle confirmed an important prediction made by TCE analyuis. One reason why the risk of
contracting out a prison service is excessive is where the contractee (in this case the Department) has to
maintain large surplus production capacity to guarantee continuity of supply should the contractor fail to deliver
according to the contract or enters bankruptcy. Unfortunately, for ERA ‘NZ Corrections boss Ray Smith said the
department had enlisted a “crack team” of 20 corrections staff that would take over the running of the prison
indefinitely.’(Serco under fire at Mt Eden Prison after NZ smartphone fight club scandal.
http://www.news.com.au/technology/gadgets/mobile-phones/serco-under-fire-at-mt-eden-prison-after-nzsmartphone-fight-club-scandal/news-story/720ee3acdca357a6506ce2d1abd2dc32).
9
Training Organisation owned by Serco. ‘Serco became so consumed with moving as many
prison guards as it could through its training school that it left its new employees at risk
when they started on the job, a New Zealand Qualifications Authority report has found. In
a report critical of Serco NZ Training, NZQA found the company's initial training course
had been deliberately structured to tick the legal boxes needed to get new employees into
prison and working. In doing so it created patchwork training leaving "Serco employees
with the basic requirements to perform as prison officers" but creating "a risk to the
employees working in a complex and high-risk environment like prisons". The NZQA
report studied the private Serco Training facility which exists solely to train prison guards
to work in Mt Eden and Wiri prisons. Serco Training is owned by Serco, the company
which has the contract to manage the private prisons’ (Serco training left new staff at risk
Nov 9, 2015 m.nzherald.co.nz).
2.3 Competition, contracting-out and innovation
We now turn to ERA’s claim that contracting-out and competition promotes innovation in
prison services. This is based on the assumption of a necessary link between ‘the profit
motive of the private sector and the role of innovation in assisting private enterprise to be
efficient, improve products, and open new markets’ ERA (2015: 247). ERA fails to note the
complex and, often, perverse effects of economic incentives on agent behaviour. The
literature on the economics of innovation clearly establishes that the relationship between
private profit and innovation is highly complex and that it is illicit to assume that competition
by itself will promote private investment in innovation. The construction industry serves as
an excellent case study. It is the second largest industry in Australia in terms of output, and
has arguably the most intensely competitive industry structure, with literally tens of
thousands of competing contractors, low barriers to entry and a mostly unregulated
workplace. Of all industries it is one of the closest to meeting the orthodox economists’
nirvana of perfect competition. Unfortunately, it also has one of the lowest rates of innovation
and productivity growth. Loosemore, Martin 2015
https://www.be.unsw.edu.au/content/opinion-why-construction-industry-needs-innovate.
Second, the innovation literature teaches us that the private sector (with the exception of
mega-corporations like IBM and ‘big pharma’) will not undertake ‘basic research’, that is,
research designed to create new knowledge. Rather, the profit motive skews private research
and innovation into applied research which uses existing knowledge to adapt products and
services to the market. By contrast basic research is inherently risky as the successful search
for new knowledge, by definition, cannot be guaranteed. Equally, there is no prospect for
immediate financial return as the findings from basic research require additional investment
and research to convert new knowledge into practical outcomes. Finally, the findings of basic
research typically cannot be patented. Put simply, the outcome of basic research is in the
nature of a ‘discovery’, like genes for breast cancer or mathematical theorems. These are
distinguished from ‘inventions’ or the output of applied research, which can be subject to
patents or copyright (Calvert and Martin 2001). The consolidated impact of all these
characteristics is that firms and competitors cannot be excluded from using the basic research
produced by others, greatly reducing the incentive of private firms to invest in this essential
activity. The significance of this extended and arcane disquisition for the inquiry into
contracting-out human services is that it is most unlikely private contractors will undertake
the type of basic research that the prison system needs. An example is investment in high
quality, large scale development and evaluation of prison programmes to improve their
efficiency and effectiveness. Indeed, confirming this analysis the WA Department of
10
Corrective Services Commissioner recently committed to invest in exactly this type of
research capacity.8
2.4 How will competition work in practice?
Finally, it is also important to highlight very briefly a few problems arising from ERAs
failure to detail the practical mechanics of its proposed tendering and competition process:
 how will efficiency and cost saving gains arise from contracting and competition? This
somewhat startling question arises from the fact that ERA does not reconcile a
fundamental conflict in its approach. On the one hand all prison services, the outcomes of
those services and the ‘efficient price’ of each service will be precisely defined by DoCS
using the methodology prescribed by ERA. This data is the basis for measuring system
performance. ERA describes this methodology in strict economic terms as providing a
precise ‘marginal opportunity cost’ of these services. Where, therefore, is the scope for
bidders to compete on price when the most efficient set of prices has already been
determined by DoCS? (The concept of ‘efficient price’ is examined in section 3).
 mandating and prescribing in enormous detail the services to be delivered and their
outcomes precludes any alternative approach to achieving desired prison outcomes.
Where, therefore, is the scope for bidders to compete on innovation when the exact set of
services to be delivered has already been determined by DoCS? If price and service is
prescribed and mandated what then is the basis for competition within ERA’s model?9
 ERA is keen for not-for profit entities to be included in competition for providing prison
services (ERA 2015: 272). This raises the interesting dilemma, by definition a not-for
profit entity is not ‘commercial’ or profit driven. The whole basis of ERAs claim that
efficiency and innovation will be enhanced by competition is that entities such as SERCO
are driven by their ‘commercial interest’. This is not an argument against the involvement
of the not-for-profit sector, rather it highlights the difficulty of incorporating involvement
of this sector into the economic theory of organisational performance proposed by ERA
 ERA does not confront the problem that the private market for provision of prison services
will be a concentrated oligopoly. Indeed ERA (2015: 257) recognises the possibility ‘that
a tender process attracts only one alternative provider (as was the case when the Wandoo
contract was tendered in 2011)’, but tries to discount the possibility of a monopoly
supplier by suggesting ‘that the presence of a public sector competitor introduces genuine
competition’. Aside from the fact that just two potential bidders is a highly concentrated
‘I have set an ambitious goal for the Department of Corrective Services to reduce recidivism rates by six
percent a year. To achieve this we must become a learning organisation. We must identify the characteristics of
effective practice and use this knowledge to pursue a culture of continual improvement. In support of this goal I
have embarked on an ambitious reform program that aims to radically change the way we do business. The
Department is now working to build the evidence base needed to design and deliver programs that are tailored
to, and proven to be effective for, different offender groups and offence types. This evidence has already told us
the programs need to be delivered within an integrated, individualised, case management framework that
extends beyond the structural confines of the Department and provides greater support for offenders upon
release. This is a role that the Department can only undertake in partnership with the community and with the
involvement of all sectors’ (DoCS 2014: Foreward).
9
ERA may respond that it avoids these problems by only prescribing an output or outcome, not the inputs or
precise activities that produce an outcome. This is consistent with the neoliberal approach to contracting. This is
meant to make space for innovation in prison service delivery. But in seeming to solve one problem, that is to
avoid prescribing ‘how’ a service is delivered, it has created another devastating problem. It is simply
impossible to establish the ‘efficient price’ of delivering a prison service without precisely detailing all of the
inputs that enter into its production. In other words, it is simply not possible to set the price of a service or
products without knowing all of its inputs. This is taken up in section 4 dealing with benchmarks.
8
11
oligopoly, it assumes that a public supplier will actually bid, or indeed be permitted to bid.
Orthodox economics recognises that in such concentrated markets the degree of
competition over price, quality and innovation cannot be predicted. In such markets firms’
decisions over price and output are interdependent and indeterminate and rely on secondguessing the reaction of a limited number of competitors. In other words, in such markets
the degree of competition depends on strategy, which is why the most common approach
to the study of oligopoly is a field of economics called game-theory (Varoufakis1995).
ERA frequently refers to the current Acacia contract as an exemplar of the benefits of
contracting-out and competition. ERA creates the impression that the contractor is given a
global fixed sum per year to run the facility and then it is in the commercial interest of the
contractor to minimise costs and drive a wedge between costs and income- by reducing the
former it increases profit. A closer examination of the contract reveals that the contract
imposes few constraints on the contractor to limit its costs, at least for several major items of
expenditure. This arises because the contract is essentially a cost-plus contract, which means
that the contactor is fully compensated for any increases in costs occurring over given period.
For example, Schedule 1 of the Acacia Prison Services Agreement between DoCS and Serco
allows for full compensation for any increase in wages and insurance above the CPI and an
automatic CPI increase is made to the Base Operating Sum, an amount designed to meet all
variable costs (DoCS 2006, Schedule 1: 120-123). Maintenance at Acacia is the subject of a
separate contract between DoCS and specialist provider.
A final and crucial observation is that despite making the virtues of competition a core
argument for contracting-out (along with benchmarking) ERA admits that due to a number of
problems simply benchmarking prisons may be a preferable and more feasible approach to
prison improvement than open tendering all prisons and services. The reasons for this are the
high cost in bidding for contracts and the anticipated slow pace of the contracting-out process
given the large resources involved in assessing bids, negotiating and customising contracts
for each prison.
Thus ERA (2015: 257) states there are ‘considerable barriers to entry in compiling tender
documents to operate a prison or supply prison services. The ERA has been told that it can
cost over $1 million to compile a tender document to bid to operate a prison. Such barriers to
entry may result in only large private sector providers participating in a tender process
(commissioning or direct procurement)’. In addition, ‘ERA observes that of the 120 prisons
under the authority of the National Offender Management Service in England and Wales,
only ten per cent of them have been subjected to a commissioning process in the six years
since a commissioning framework was introduced in 2009. If the same proportion of prisons
in Western Australia were subject to commissioning over the proposed five year term of a
Service Level Agreement, only one or two would be subject to a commissioning process
(ERA 2015: 264). These constraints lead ERA (2015: 258) to conclude it ‘is neither practical
nor optimal for the Department to subject every prison to a commissioning process at the
expiry of its Service Level Agreement. The cost alone would be prohibitive when compared
to the likely benefits’.10 This last statement is truly extraordinary, if the cost of subjecting all
prisons to successive rounds of competition (commissioning) exceeds the likely benefit,
where is the justification for subjecting any prison to competition?
In a footnote to this statement ERA (2015: 258) expands on this claim. ‘The cost of undertaking a
commissioning process for the Department is considerable. Given the considerable cost to non-public providers
of participating in a commissioning exercise, it is unclear whether they would have the financial appetite or
human capital to participate in multiple commissioning exercises concurrently.’
10
12
3. Impossibility of precise cost comparisons and cost-benefit analysis in heterogeneous
human services
Over a period of more than 2 years ERA conducted an exhaustive inquiry into the case for
contracting-out WA prisons and introducing competition between prisons through the process
of ‘commissioning’, where existing public and private providers bid for funding based on
meeting financial and operational performance benchmarks ERA could not finally determine
first, whether there would be efficiency gains from its radical proposals and second, the
quantum of these gains. ‘The ERA acknowledges that it has not costed the introduction of
commissioning. Conducting a full cost-benefit analysis of reform is a significant undertaking
and was not possible with the time-frame and resourcing provided for this Inquiry’ (ERA
2015: 239). This is an unsound basis for public policy. The critical point is that ERA’s
advocacy for contracting-out was based on a priori reasoning founded on neoclassical
principles and was not based on evidence of relative costs and benefits of public vs. private
provision of prisons. The reason for the absence of evidence is that the precision in costing
and pricing of inputs and outputs across a service delivery organisation as complex as a
prison system is impossible. Even achieving approximate costings requires a heavy reliance
on assumptions, many of which are highly disputed. This section details some of these
reasons for the impossibility of arriving at precise, or even approximate, comparisons of costs
and outcomes in complex services delivery organisations.
Competition between and within public and private providers and benchmarks specified in
contracts are the keys to driving improved performance. ‘The ERA identified four categories
of prison performance that should be measured in benchmarking. These categories are safety
and security, rehabilitation, prisoner quality of life and management. Each category
represents an outcome that a good prison should deliver’ (ERA 2015:143). These four
objectives have their associated 26 performance measures. ERA (2015: 125) approvingly
cites the UK benchmarking model developed by the National Offender Management Service
as the template for its approach. ‘The purpose of the program was to:
 Define what should be delivered (the outcomes and outputs for services) – so consistent
services could be delivered across all areas of the business;
 To know exactly what it costs to deliver each service’ (ERA 2015: 125)
 In addition ERA proposes it own method to compare performance across prisons by
controlling each benchmark for factors such as the range of different prisons and prison
classifications and prisoner needs.
It will be shown that ERA’s proposed specification, costing and benchmarking programme is
subject to four distinct and insoluble problems. First, how to establish a causal relationship
between a particular prison service or activity and a desired outcome; second, the ERAs
concept of an ‘efficient cost’ requires calculating the ‘marginal opportunity cost of a service’,
and thirdly this cost must be adjusted for all factors that can affect both the price and outcome
of all services across different prisons (ERA 2015: 134). Finally, there is a fundamental
contradiction between ERAs demand for precision in specifying standards, outcomes and
prices of prison services but also requiring contractors to innovate.
3.1 Establishing a causal relation between services and outcomes
In constructing prison benchmarks ERA explicitly states it is only interested in ends and not
means: only prescribing outcomes that prison services are supposed to deliver. Contracts
13
should not prescribe the inputs or processes to achieve these outcomes. It does this because
indifference on the part of the principal to means creates room contractors to innovate in
achieving outcomes with potential gains in quality and lower cost. Specifying only outcomes
in a contract between DOCS and a contractor is also seen as the best way to ensure clarity
and accountability over what precisely is being bought with taxpayers’ money by the
Department and to achieve value for money.
There are three basic problems with this approach to benchmarks. First, ERA cannot be and
actually is not indifferent to how outcomes are achieved. Second, it is not true that the price
of a good or service can be given without precisely specifying all the inputs that go into its
production. The only way ERA can know the cost of prison services is to precisely define
what they are and the standards they are to achieve. This, in fact, is exactly what the
specification, benchmark and costing exercise is intended to do.11 The reason for this gap
between the intention, to focus only on outcomes, and the reality can be found in ERAs
adherence to the neoliberal world-view, which, as noted above, gives priority to notions of
contractual flexibility and experimentation. Gary Sturgess (2012: 8) for example writes that:
‘Contracting for outcomes rather than inputs...and paying providers based upon their results,
also serves to increase the diversity of service models, since [principals]...are not specifying
the ways in which inputs are to be connected to outputs and outcomes. Payment by-results
has increased the amount of experimentation with alternative service models, results has
increased the amount of experimentation with alternative service models’.
But these two problems are relatively minor issues, and could be dismissed by some as mere
academic quibbling. The third problem however, is far more serious. This problem was
highlighted in the Western Australian Prison Officers Union’s response to the ERA Draft
Report and is neatly summarised by ERA (2015: 135) in their Final Report:
‘Dr Toner submits that to be of analytical use, the ERA performance measures and
benchmarks have to solve three quantitative problems:
precisely define the prison services that contribute to achieving the four objectives
of the prison system;
rigorously establish the relationship between the quantity and price of each prison
service and the extent to which it contributes to achieving prison objectives; and
complete these tasks for all sixteen Western Australian prisons...
That ERA’s specification, benchmark and costing exercise is actually focussed on inputs is evident from its
description of how performance in meeting the core objective of prisoner rehabilitation is to be measured. What
is being measured is not the outcome itself but simply the chief inputs or services that are presumed to cause this
outcome. ‘The performance of individual prisons in rehabilitating prisoners is difficult to measure...prison
Superintendents have little control over many of the factors that contribute to a prisoner reoffending...There are
seven factors that are identified to contribute to the likelihood that a prisoner will reoffend on release. These
factors are: accommodation; education, employment and training; budgeting and debt management; drug and
alcohol dependence; physical and mental health; family connection; and attitudes, thinking and behaviour.
Therefore, the performance measures the ERA has recommended in this area are largely limited to a prison’s
ability to maximise prisoner participation in, and completion of, programs, education and employment (ERA
2015: 144-145). An even larger problem arises in measuring the other core prison objective ‘prisoner quality of
life’. ERA simply sets out a number of options, including surveys and a particularly resource intensive method
called MPQL that it notes but does not recommend any specific measure (ERA 2015: 145).
11
14
‘He] considers that even if it were possible to develop such benchmarks, it would only be
warranted if two conditions are met: the Department must be able to fully control all inputs of
the prison system; and the Department must also fully control expected outputs (such as
improvement rates in recidivism and rates of efficiency improvement)... Dr Toner submits
that neither of these conditions apply...[and] concludes that under these circumstances,
benchmarks and performance standards act not to drive improvements, but simply to quantify
the gap between expectations and reality. Dr Toner states that performance targets and
benchmarks are important for organisations...[and] considers this is true if the setting of these
targets and the data collection are both feasible and useful. Dr Toner considers the sheer
complexity and unrealistic ambition of the ERA’s proposal means they are neither feasible
nor useful’.
ERA does not dispute this characterisation of what it is recommending. Indeed, the unlikely
response was that ‘ERA agrees with Dr Toner that the Department needs to undergo a
process of defining the services delivered in prisons that contribute to achieving objectives
and the price and quantity of those services in each prison’ (ERA 2015: 136). In other words,
ERA casually suggests that more work needs to be done without confronting the mindboggling complexity of what it is recommending, nor does it give any clarity about the
method for the task. This all left for DoCS to solve.
The issue here is not specifying prison services and the delivery standards, as noted
previously, that task has already been undertaken by DoCS in exhaustive detail in its Policy
Directives, Adult Custodial Rules, Assistant Commissioner Custodial Operations Notices,
and Compliance Testing Standards for public prisons. Rather, the problem lies in ERA’s
insistence that DoCS must devise a method that will establish a causal relationship between a
precisely defined quantity and quality of services and precisely defined outcomes. ‘ERA has
been mindful that cost allocation should be based on principles of causality, objectivity,
consistency, and transparency’ (ERA 2015: 53). Unless this relationship can be established it
is not possible to show that the services and benchmarks recommended by ERA are an
improvement on the current approach. Second, it must establish that prison services and
standards are the uniquely optimal solution in terms of efficiency and effectiveness in
meeting prison objectives. Unless this second problem is solved the claim that ERAs
benchmarking and contracting model will deliver more efficient and superior services is
undermined. (This second problem is taken up in section 3.2).
However, the theoretical and practical problems in establishing the relation of cause and
effect in prison services and outcomes are prodigious. To take an obvious example, exactly
what services contribute to achieving the objective of safety and security within prison?
Possibly, all of them? (More formally, the problem that arises here is the pricing of ‘joint
products’ where inputs from a production process lead to multiple outputs). It is most
unlikely that services say, directed at the objective of rehabilitation will not positively affect
the behaviour of prisoners whilst still incarnated. If so, it raises the difficult issue of how
then, to allocate the benefit and cost of services directed to the rehabilitation objective to the
security objective? If this is not done, inefficiencies will arise as services are being paid for
but the benefits and effect on achieving objectives are being incorrectly assessed. This will
result in services being either over-supplied or under-supplied. Inability to accurately allocate
the costs and benefits of particular services across objectives will be an especially pressing
problem where, under contracting-out, different prison services are divided by different
private firms or even different divisions within the same firm. Once these ‘spill-over’ effects
become apparent, say in the example above, this will lead a supplier of security services to
15
diminish supply of their service and rely on a different supplier to maintain their
rehabilitation service.
ERA concedes the obvious practical problem that, given the current state of knowledge and
expertise, DoCS is just unable to answer these questions. Moreover, even assuming the
theoretical problems can be addressed it will take many years and probably decades for
rigorous information to be collected to construct service specification, benchmarks and
costings. Contracts require DoCS to ‘clearly define the services to be delivered, and the
funding and staffing required to efficiently deliver those services. This is fundamental to
ensuring that prisons have sufficient financial and staff resources to deliver the service
standards expected under Service Level Agreements ’(ERA 2015: 265). ‘However...the
Department faces a number of longstanding issues with its planning and processes, and in the
way in which it collects and uses information to make good decisions’ (ERA 2015: 121). In
particular, the Department lacks analytical and research capacity in programme evaluation.
‘In order to deliver programs effectively, the Department needs to:
assess how effective those programs have been for individuals; and
fective programs have been at addressing the
needs of prisoners collectively’ (ERA 2015: 213). This is an important and welcome
recommendation, though ERA does not acknowledge that rigorous evaluations can take a
long time.
The ‘gold standard’ for programme evaluation is to randomly allocate matched cohorts of
prisoners to different interventions with another matched ‘control’ group receiving no
interventions and then assess the outcomes. In the case of assessing the effect of programmes
on say, recidivism this is not only extremely complex where prisoners undertake multiple
programmes but also can take years for outcomes to emerge in the case of long-term
prisoners but also sufficient time needs to lapse after release from prison to properly assess
the outcome. ERA already accepts that it is not possible to evaluate the actual rehabilitation
outcomes of a single prison as prisoners typically pass through multiple establishments over
the course of their stay.
3.2 What is an ‘efficient cost’?
Despite the great weight ERA gives to the concept of an ‘efficient cost’ in its model to create
a more efficient and effective system the term is never explicitly defined. Given the term is
used by an entity called the Economic Regulation Authority; it is not unreasonable to suggest
the term has an orthodox economic meaning. This is also supported by the various
connotations given by ERA to the term as well as its published statements on the definition of
economic efficiency. 12 We use these various sources to build up a meaning of the term:
First, central to the orthodox idea of efficiency is equating the marginal cost of a commodity
c, that is, the cost of production of the ithunit of output of c with the jthunit of revenue earned
12
For participants in the WA wholesale electricity market ERA produced a paper defining the economic
meaning of short and long-run marginal price. Apparently, there was some confusion in this market due to
different meanings of these terms viewed as accounting concepts or economic concepts. ERA (2008) Portfolio
Short Run Marginal Cost of Electricity Supply in Half Hour Trading Intervals Technical Paper (Author: Adam
McHugh)
https://www.erawa.com.au/cproot/6317/2/20080111%20Short%20Run%20Marginal%20Cost%20%20Technical%20Paper.pdf
16
on that output. Maximum allocative efficiency is achieved when marginal cost and marginal
revenue equal the price of the commodity set under perfectly competitive conditions. This is
to be distinguished from average cost which is just total output divided by total revenue.
ERA (2015: 123) explicitly supports the use of marginal analysis for allocating resources
efficiently in the WA prison system: ‘the contract for Acacia sets out the funding due to
Serco for different levels of prisoner population. This approach acknowledges that there is a
marginal cost to increasing a prison’s population. The ERA recommends that Service Level
Agreements with public prisons contain similar funding tables’.13
Maximum allocative efficiency requires markets to be ‘perfectly competitive’, though, the
actual market for prisons is and will remain be oligopolistic.14 ERA seems content not to
investigate this thorny issue in any detail, as to investigate too deeply would reveal the
essential indeterminacy of outcomes from markets which are not ‘perfectly competitive’.
ERA certainly anticipates significant advantages in terms of cost and efficiency from
adopting its recommendations to enhance ‘competition’.
Second, as ERA allows that its recommendations will entail adjustments in all inputs,
interventions and even constructing new more suitable prisons, the ‘efficient cost’ of
delivering these changes is thus a ‘long-run’ marginal cost.
Third, as all consumers are resource constrained they must choose what commodities to buy
or forgo. The inevitability of choice is known as ‘opportunity cost’- to gain one commodity
another ‘must be given up’ (ERA 201: 129). Optimising these choices involves equating the
marginal utility (or benefit) of a commodity with its marginal price and comparing this with
the ratio of utility to price of all other commodities.
Four, consumers must have some way of adjusting the prices of similar commodities for
quality differences and similarly, producers must have some way of adjusting the prices of
similar inputs to production that they buy to make commodities. An efficient cost requires the
construction of quality adjusted prices for inputs and outputs. Orthodox economic theory
assumes that consumers create valid and reliable quality-adjusted price indexes of the output
of firms and producers quality-adjusted price indexes of their inputs used in production.
Optimising marginal decision-making is still required; it is simply augmented or complicated
by the need to adjust these decisions for quality differences. Applied to the analysis of prison
reform ERA (2015: 48) argues that a ‘substantial, but well-targeted investment that raises
current costs, can be efficient in the long-term, reducing recidivism, and consequently
avoiding the future costs for the Department, courts, police, and the wider community’. In
other words, an ‘efficient cost’ for prison services is not necessarily the lowest because of
quality differences between services producing different outcomes.
13
Actually, the contrat between DoCs and Serco for running Acacia is not based on marginal cost adjustments
for changes in prison population. ‘In 2013-14, on average, 3,220 prisoners had served a prior prison sentence.
With a cost per prisoner per day of $334 (in 2013-14), this equates to a daily cost of $1,075,480. This figure is a
reflection of total cost. While the ERA acknowledges that a figure based on marginal cost would be more
accurate, it has been unable to calculate this cost because the Department has been unable to provide the
requisite data (ERA 2105: 73). Apparently, DoCS have been charged by ERA with supplying this data.
14
One of these conditions that definitely informs ERA’s approach is the assumption that all ‘economic agents’
have perfect knowledge of all possible influences on price and output. Only with such an assumption can ERA
recommend DoCS compile information that, in effect, gives it and all competitors the necessary information to
optimise all decisions.
17
Finally, a variation on this issue is that comparing the efficiency of prisons in meeting their
core objectives and performance benchmarks requires adjustment for differences in the
capacity of prisoners and prisons to convert inputs into outputs. For example achieving the
same performance on security will be more costly for a maximum security prison to achieve
than a low security prison farm. Thus ‘two prisons performing equally effectively will not
necessarily have equal costs’ (ERA 2015: 48).15
Efficient pricing is thus ‘the long‐run marginal opportunity cost’. Applied to the prison
system in making such key decisions as what rehabilitation services should be purchased it
requires the criteria of price, quality and effectiveness of one intervention be precisely
compared using the same criteria against all other possible interventions. 16 Another way to
express this is that an ‘efficient price’ is actually a way of measuring change in the
productivity of prison services and prisons.
Lest it be thought that the foregoing represents a caricature, ERA applies this logic in the
analysis of rehabilitation outcomes from prison industries. WA prisons produce a whole
range of commodities in prison industries such as ‘primary produce, catering, laundry and
clothing’ as these ‘can reduce the cost of Imprisonment’. More importantly, employment in
prison industries also generates rehabilitation outcomes. However, from ERA’s perspective
this creates a resource allocation problem as ‘each prison industry will have an opportunity
cost in terms of time and money... prison industries should not be pursued simply because
they may generate some rehabilitation outcomes. Industries should only be pursued if they
will result in the best rehabilitation outcomes for a particular level of investment in time and
money. If not, the Department should consider investing in other, non-industry, rehabilitation
activities’ (ERA 2015: 129-130 italics added). ERA (2015: 275) concluded with a formal
recommendation to establish the efficient cost of these services: ‘the Department of
Corrective Services undertake a rigorous cost-benefit analysis of prison industries. The
Department should only continue these industries where it can be demonstrated that the
benefits outweigh the costs and the net benefits exceed those of alternative activities’. 17
One factor that drives the need to establish precise cost comparisons for ‘similar’ services is that in a large
state like WA there will be large differences in the rate of inflation of these services, not just differences in price
levels which ERA notes (2015: 65).
16
This view of what ERA is means by the term ‘efficient cost’ is identical to that employed in the UK by
specialists at the London School of Economics who were commissioned by NOMS to efficiently price prison
services in preparation for commissioning. ‘The costing approach was guided by economic theory, reflecting the
opportunity cost of activities undertaken...The aim has been as far as possible to measure the opportunity cost of
the activity or intervention to society in order that the full resource implications of opportunities or benefits lost
are reflected, rather than expenditure to a single agency or sector. In order to reflect the full resource
implications of any activity, it is also important to identify the resource consequences in the long‐run...the focus
is on the impact of implementing changes on the margin of the offender management service as a whole. A
marginal cost is the cost of an additional unit of activity or providing a service to one extra person. Based on
these economic principles, the aim has been to estimate (as far as possible) the long‐run marginal opportunity
cost to society’’ (Nadia Brookes, Barbara Barrett, Ann Netten and Emily Knapp (2013) Unit Costs in Criminal
Justice (UCCJ) http://www.pssru.ac.uk/archive/pdf/dp2855.pdf p.7).
15
Needless to say, as with ERA, this grand ambition, guided by orthodox economic theory, could not be realised.
What was actually produced by the LSE group was a remarkably thorough and detailed adding-up exercise of
the unit costs of a great number of prison services, mostly focussed on rehabilitation and health. There was no
comparison of the quality-adjusted price of services controlled for their relative effectiveness in achieving their
objectives.
17
ERA demands DoCS solve an even more complex optimisation problem when it comes to self-sufficiency
and rehabilitation outcomes for prison industries. Many prison industries supply all the needs of the prison in a
18
The fictional character of an ‘efficient cost’ is also revealed in ERA’s admission that it
cannot calculate something as basic as an appropriate ‘prison capacity... rate [or] the daily
average prisoner population as a percentage of the capacity of the prison’ (ERA 2015: 117).
This is ‘because there are multiple definitions of prison capacity and competing views among
corrective service agencies and stakeholders as to which of these definitions should be used
to measure prison utilisation’ (ERA 2015: 117-118). This is an extraordinary admission as
without a rigorous scientific basis for the optimum number of prisoners within a facility it is
impossible to know the efficient level of resources a prison should receive. ERA (2015: 180)
notes this is a serious problem in its attempts to identify ‘a prison’s capacity to provide
services and its staffing levels. Prisoners require access to services such as health, education
and rehabilitation programs in order for a prison to achieve its objectives. Similarly, prisons
require a sufficient number of prison officers in order to safely accommodate its prisoners....
The ERA identified rehabilitation and health as the services that should be considered in a
measure of capacity. The ERA sought advice from key stakeholders on how the capacity of
these services could be measured. Stakeholders expressed a view that measuring service
capacity is complex because service needs vary significantly between prisoners cohorts...
There is similar difficulty in assessing required staffing capacity. The number of staff
required in a prison differs based on the difficulty of managing particular cohorts of
prisoners... Therefore, due to the complexities in measuring services capacity and required
staffing levels, the ERA has not developed an alternative measure of prison utilisation.’ ERA
adopts a pragmatic solution and adopts ‘design capacity’ as its preferred measure of prison
capacity. Now, this may well be a desirable outcome for the welfare of prisoners but it is not
based on economic principles and therefore cannot be deemed an efficient outcome in the
sense that ERA would use that term.
To conclude this discussion, not only are the information requirements to derive ‘efficient
costs’ huge, more importantly, the data relating to say, relative rehabilitation outcomes from
all actual and possible prison activities, simply does not exist. As argued above it could never
be obtained. The essential point for policy is that unless these exacting, and indeed,
impossible, conditions are meet there is no a priori argument to justify ERAs approach to
prison improvement over any other approach. Second, because of the multiple problems in
actually measuring ‘efficient cost’ ERA simply cannot empirically, demonstrate that its
specification, benchmarking and costing programme is superior to alternative approaches to
improve system performance. Finally, it shows the limits of using the ‘other-worldly’
assumptions of orthodox economics, especially in its extreme form, for public policy.
(Another major flaw in the idea of efficient cost is revealed in the next section).
range of commodities and their production also leads to certain rehabilitation outcomes. For all of the
commodities in which the system is self-sufficient DoCS is required to assess not only whether it is cheaper to
source these from external suppliers but also to assess the rehabilitation benefits derived from these selfsufficient industries against all alternative rehabilitation interventions. ‘The ERA acknowledges that it may be
possible to justify paying a higher price for goods and services produced through prison industries (compared to
the price of purchase from external providers) to reflect the rehabilitation benefits achieved for prisoners.
However, again, the Department has not undertaken sufficient analysis to determine the extent of rehabilitation
benefits being achieved from prison industries focussed on self-sufficiency and cost of these benefits’. It is also
worth noting that the achievement of self-sufficiency in the production of certain items really ‘grates the gears’
of orthodox economists. Self-sufficiency is equated with the ultimate economic ‘sin’: industry protection and
‘autarky’ prior to nations’ entering the blissful state of ‘free trade’.
19
3.3 Problems in Comparing Prison Cost and Performance
Section 2.2.3 described in detail ERA’s proposed approach to use its specification,
benchmarking and costing programme to enable valid comparisons across prisons. Aside
from the obvious comparisons of performance against benchmarks, such comparisons are
essential to ensure equity in funding across prisons. To make valid comparisons a number of
problems must be solved. The first is how to compare the cost of prisons? ‘Costs and service
requirements will vary from prison to prison, reflecting the specific circumstances of each
prison. Service Level Agreements for individual prisons will need to reflect this variation to
ensure an appropriate level of funding’ (ERA 2015: 125) Second, is it possible to validly
compare the achievement of different prisons against key performance indicators when it is
known that objectives like rehabilitation are strongly affected by factors such as the
classification of prisoners? For example, maximum security prisoners will be more resistant
to the effects of rehabilitation programmes than minimum security prisoners.
This section identifies difficulties with ERA’s solutions. First, it is not possible to reconcile
the definition of an ‘efficient price’ and the central place it has in ERA’s analysis with its
claim that it is not possible to make valid comparisons of cost per prisoner across different
prisons. Second, there are major issues with ERA’s selection of the independent variables to
adjust for differences in the capacity of prisons to achieve benchmarks.
To briefly recap ERA’s analysis of prison benchmarks. ERA says the first problem identified
above is insoluble as it agrees with WAPOU’s submission to the Draft Report that it is not
possible to adjust for all the factors that lead to cost differences across prisons. ‘Dr Toner
considers that it is not possible to develop cost benchmarks because of the high variability in
cost per prisoner per day, driven by differences in the prison population, prison design and
age, and prison location. He considers that it is not possible to directly compare prison
performance without adjusting for these factors...The ERA agrees with submitters that cost
per prisoner per day is an inappropriate measure of prison performance, particularly in
Western Australia where few, if any, prisons are directly comparable... this measure has been
removed from the ERA’s list of recommended performance measures’ (ERA 2015: 140141).18
Despite this limitation ‘ERA does consider that per prisoner per day (on both a system wide
and individual prison basis) provides important information to both management and
stakeholders as to cost drivers, and the outcomes of investment and policy decisions’ (ERA
2015: 141). ERA classifies ‘cost per prisoner per day’ as ‘management information’ but
retains the other three core objectives (safety and security; rehabilitation; and prisoner
quality of life) as benchmarks with their associated performance measures. Thus the prison
management category is ‘designed to measure productivity in prisons [and] the other three
categories are designed to measure quality of service’ (ERA 2015: 141).
The second problem is solved by applying four independent variables which ERA claims
account for the bulk of variation in the capacity of different prisons to achieve performance
‘Cost per prisoner per day simply describes the Department’s use of funds. It provides no information on the
appropriateness of the use of those funds. For example, a prison with a very low cost per prisoner per day but
very poor rehabilitation outcomes may be inefficient, but a prison with a high cost per prisoner and very low reoffending rates may be using those funds well to deliver an efficient outcome for the State’ (ERA 2015: 54).
18
20
benchmarks. 19 ‘Under this approach, the ERA has proposed that benchmark targets be
adjusted to reflect differences in the composition of each prison’s population. In particular,
the ERA has proposed that the Department adjust benchmark targets to reflect differences in
the security classification (that is, maximum, medium and minimum-security), sentence
status (that is, remand or sentenced) and gender of the populations of individual prisons
(ERA 2015: 134).
3.3.1 Cost per prisoner per day
Admitting that is not possible to derive a method for accurately comparing per prisoner per
day cost across prisons is inconsistent with the demand that DoCS calculate an ‘efficient
cost’. ERA’s design specification for what an efficient cost is meant to measure logically
contains the means to make valid cost comparisons across prisons.
In response to the WAPOU submission to the Draft Report ERA (2015: 134) concedes that
benchmarking requires all of the following be solved and quantified:
prisons and the wider prison system... Dr Toner further submits that benchmarks must also
control for differences in prison characteristics that cause differences in the type of service,
their cost and achievement of the objectives set for individual prisons and the wider prison
system...The ERA considers, based on consultations with the Department, that the
recommended approach to benchmarking is both feasible and useful. The ERA agrees with
Dr Toner that the Department needs to undergo a process of defining the services delivered in
prisons that contribute to achieving objectives and the price and quantity of those services in
each prison’ (ERA 2015: 134 italics added). In theory this data is sufficient in itself to enable
cost comparisons across prisons. Demanding the use of ‘efficient cost’ but denying the
comparability of costs across prisons introduces a fundamental ambiguity into ERA’s
analysis and recommendations.
Expressed differently, if it is not possible to adjust for all the factors that give rise to cost
differences across prisons what is being measured is a not an ‘efficient cost’ it is simply cost.
The whole idea of calculating an efficient cost is that it permits not only accounting for
changes within a prison that lead to cost changes, say in response to a change in the
classification of prisoners or the introduction of new technology, but critically to explain why
there has been a change in costs. Remember, an ’efficient cost’ is not a bookkeeping identity;
it is an analytical device to measure the long-run marginal opportunity cost of an individual
service, and all services combined to optimise performance measures. In addition, it
explicitly entails controlling for all the variables within a prison, such as those identified
above, that lead to cost variations in meeting performance targets. This creates the basis for
valid cost comparisons across prisons.
Secondly, ERA is inconsistent in its statements about this issue. ‘The Department does not
have robust information on how much it currently costs individual prisons to deliver specific
19
Applying weights to these performance measures to control for the different priorities DoCS may have for
different prisons is just a complexity to the calculations and is not dealt with here.
21
prison services or the efficient cost of delivering those services. Developing this
understanding is a foundation step in ensuring that Service Level Agreements are realistic
and achievable. Reflecting this, the ERA recommends that the Department engage in a
specification, costing and benchmarking program (as detailed in Box 4)’ (ERA 2015: 124
italics added). One of the benefits identified from introducing a specification, costing and
benchmarking program in Box 4 is that ‘access to robust cost information allows the
Department to compare the cost of individual prisons and better compare the cost of
operating the more expensive (and presumably older) prisons with the cost of replacing them
with new facilities that have lower operating costs‘(ERA 2015: 125 italics added).’ This
introduces a fundamental ambiguity into ERA’s analysis and recommendations.
Thirdly, having ‘solved’ the problem that in reality it is not possible to compare cost per
prisoner per day across prisons ERA has created an even larger problem. Without this basic
metric there is no method for DoCS to rationally and equitably allocate resources across
prisons. On the one hand, ERA says it can compare the performance of prisons through its
benchmarking and target system, but it is impossible to compare how efficiently resources are
used across prisons to produce these outcomes since it is invalid to compare the costs of
running prisons. There is therefore, no system to allocate resources across prisons on the
basis of efficiency. ERA’s edifice of specification, benchmarking and costings, SLA’s and
contracting-out cannot therefore, achieve its central goal of lifting efficiency and
productivity, since they cannot be measured.
Finally, an obvious critical problem with denying that cost variations across prisons are
commensurable is that there is no basis for comparing the efficiency, or more specifically the
rate of productivity change, of different prisons.20 This latter issue creates numerous
problems such as how is DoCS to set equitable financial incentives for Superintendents to
reward improved productivity within and across a prison when there is no basis for
comparing the rate of change of productivity. It may well be the case that it is easier for one
prison to improve its productivity compared to other prisons. In fact, this is the whole basis
for ERA’s recommendation to adjust performance benchmarks for different characteristics of
prisons that affect the capacity of prisons to achieve these benchmarks. It is to this issue we
now turn.
3.3.2 Selection of the four independent variables to adjust benchmark targets
To adjust for the differing propensity of prisons to convert inputs into outputs or outcomes,
ERA recommends, as essential to its specification, benchmarking and costing programme,
that three control or independent variables be used to enable the valid comparison of prison
performance. ‘The main factor that is likely to affect a prison’s performance is the
composition of its population. These factors should be adjusted for in setting a prison’s
benchmark targets... the ERA considers that the factors that have the most influence on prison
performance are security level, sentence status (that is, remand or sentenced) and gender’
(ERA 2015: 147).
There are prodigious problems, identical really in nature to those applying to ERA’s demand
that DoCS establish ‘causal’ linkages between prison interventions and outcomes to allocate
20
Efficiency is measured by reductions in the quantity of inputs to achieve a given output. The quantity of
inputs is given by their volume and their price. It also requires adjusting both inputs and outputs for quality
differences.
22
funds and also controlling for all the variables entering into the calculation of an ‘efficient
cost’.
ERA provides no rigorous quantitative basis for selecting these three control variables for its
‘population-adjusted’ approach to weighting each prison’s benchmark targets. Rather they are
based on ease of calculation and some assumed statistical reasons. ERA (2015: 151) even
admits that a ‘number of other population and prison characteristics may have some effect on
prison performance. These factors include the location of the prison, the age and offence
profile of prisoners, the proportion of prisoners who are Aboriginal, and the sentence length
of prisoners. While the ERA appreciates these factors may have some effect on prison
performance against benchmark targets, it has sought to limit the complexity of the process
by restricting the factors considered to those likely to have the greatest effect.’ This ‘near
enough is good enough’ attitude to the allocation of scarce public resources simply ‘not good
enough’.
This criticism is not, however, a call for more expensive research to improve the quality of
the independent variables. This call cannot be made since, as noted previously, there is no
rigorous cost-benefit analysis concussively demonstrating the value of ERA’s approach. This
is an unsound basis for public policy.
To conclude, because of these fundamental flaws in the model proposed by ERA, it cannot be
implemented and the forecast efficiency gains cannot be achieved both for practical reasons
and simply, as they cannot be measured they cannot be established to exist.
Bad ideas are, nonetheless, tenacious. A possible scenario is the following. A government
committed to neoliberal views will be highly attracted to a policy that promises it can deliver
more for less. The fact that it cannot actually be implemented could well be part of its appeal.
Not only would its implementation be monstrously expensive but ERA raised the possibility
that an actual increase in cost per prisoner could well be justified on ‘efficiency’ grounds.
What can actually be implemented is a replica of the UK specification, benchmarking and
costing programme. Specified service standards and their ‘efficient cost’ will be equated to
minimum standards and the lowest price. This is precisely what happened in the UK. NOMS
publishes an annual Costs per place and costs per prisoner report, in which ‘cost per prisoner
is the average cost of holding one prisoner for the year. It is the Direct resource expenditure
or Overall resource expenditure divided by the average prison population’
(https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/471625/costsper-place.pdf . There is no adjustment for any independent variables that affect cost, because
as argued above, the task is simply impossible to do rigorously. This leads to stark
comparisons such as the following where NOMS extols the ‘efficiency’ gains from its latest
round of commissioning:
‘HMP Oakwood, a 1,605-place prison, will provide places at the lowest operational unit cost
in the estate at around £13,000 per prisoner per year. This compares with an average of
£27,400 per prisoner per year...this results in the lowest direct establishment unit cost in the
prison estate’ Ministry of Justice (2012) Offender Services Competitions Annual Update 2012
(https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/217274/comp
etition-strategy-offender-services-2012.pdf) 21
21
ERA tries to disassociate itself from any adverse effects on the working lives of prison staff should its policy
be actually realised. ‘Under the ERA’s recommended benchmarking approach, the ERA does not advocate for
or expect any change to the number of prison staff or the way that they undertake their jobs as a consequence of
23
References
ABC (2015 June) Serco ordered to pay $1 million in penalties over failure of Fiona Stanley
Hospital contracts, http://www.abc.net.au/news/2015-06-10/serco-ordered-to-pay-1-millionin-penalties/6536150
ABC (2015 February) Government recovers over $41 million worth of false claims after
'rorting' of Job Services Australia scheme http://www.abc.net.au/news/2015-0223/government-recovers-millions-after-rorting-of-jobs-scheme/6193022
Bita, N (2015) Wolves of Vocational Education and Training make a billion-dollar killing,
The Weekend Australian, Inquirer, 7-8 November, p. 19
Cahill, D., Edwards, L., Stilwell, F. (2012) Neoliberalism: Beyond the Free Market.
Cheltenham, UK: Edward Elgar Publishing.
Jane Calvert and Ben R. Martin (2001) Changing Conceptions of Basic Research?,
http://www.oecd.org/sti/sci-tech/2674369.pdf
Chester, Lynne (2013) The Impacts and Consequences for Low-Income Australian
Households of Rising Energy Prices,
https://retreview.dpmc.gov.au/sites/default/files/webform/submissions/06052014%20Chester
%20-%20Impacts_Consequences_Low_Income_Households_Rising_EnergyBills_Oct2013.pdf
Corrective Services WA (2009) Inquiry into the privatisation of prisons and prison-related
services - NSW Parliament,
https://www.parliament.nsw.gov.au/prod/parlment/committee.nsf/0/819e03567c89bb85ca257
5c3008338c8/$FILE/090422%20Brian%20Lawrence%20DCS%20%20WA%20Answers.pdf
(2015) Department of Corrective Services' Annual Report 2014-15
http://www.correctiveservices.wa.gov.au/_files/about-us/statistics-publications/annualreports/DCS-2014-2015/dcs-annual-report-2014-2015-online.pdf
Economic Regulation Authority (2014) Issues Paper: Inquiry into the Efficiency and
Performance of Western Australian Prisons
(2015) Inquiry into the Efficiency and Performance of Western Australian Prisons – Final
Report
(2015) Prison Benchmarking Manual
Serco (2015) Serco Group plc Annual report and accounts 2014
https://www.serco.com/Images/22000_Serco_AR14_tcm3-46422.pdf
Streeck, Wolfgang (2015) Interview: capitalism, neo-liberalism and democracy, Renewal. A
journal of social democracy, Vol. 22 No 3/4 pp.74-80
the Department collecting more information on the costs of running prisons. Operational matters of this nature
are for the Department and Superintendents to determine’ (ERA 2015: 112).
24
http://www.renewal.org.uk/articles/interview-capitalism-neo-liberalism-and-democracy
Sturgess, Gary (2012) Diversity and Contestability in the Public Service Economy
(2013) Bound for Botany Bay Contracting for Quality in Public Services, Discussion Paper
No. 1, The Serco Institute
‘’The Sources of Benefit in Prison Contracting’’, in Delivering Justice The role of the public,
private and voluntary sectors in prisons and probation, Edited by Vicki Helyar-Cardwell
The Australian (2015) WA dumps Serco prison transport contract
http://www.theaustralian.com.au/news/latest-news/wa-dumps-serco-prison-transportcontract/story-fn3dxiwe-1227400755511
Quiggin, John (2003) Word for Wednesday: managerialism (definition)
http://johnquiggin.com/2003/07/02/word-for-wednesday-managerialism-definition/
Toner, Phillip (2014) Contracting out publicly funded vocational education: A transaction
cost critique, The Economic and Labour Relations Review 4, Vol. 25(2) 222–239
Varoufakis, Yanis (1995) Game Theory: A critical introduction. London and New York:
Routledge,
25